Inside Higher Ed points me to a new study [PDF], “What to Withdraw: Print Collections Management in the Wake of Digitization,” on how academic libraries should cope with the increasing availability of scholarly journals in digital form. They go through great pains to argue the need to maintain print. From the executive summary:

This analysis finds several rationales for retaining some copies of the print version: the need to fix scanning errors; insufficient reliability of the digital provider; inadequate preservation of the digitized versions; the presence of significant quantities of important non-textual material that may be poorly represented in digital form; and campus political considerations. The appropriate disposition of print copies of a given journal should vary depending on the characteristics of the print original and its digitized version in each of these categories.

Because many of the rationales for retaining print are likely to decline over the course of time, this report introduces time horizons for print preservation. Librarians have often discussed preservation responsibilities as if it were possible to undertake perpetual commitments, but specified time commitments coupled with regular reassessment of priorities and responsibilities permit better decisionmaking. The model we propose therefore examines the minimum period of time that access will be
needed to at least one copy of the print original.

While complex, this methodology provides for preservation frameworks that vary based on risk profiles. For example, text-only materials require less concern than image-intensive materials, while high-quality digitization processes digital preservation practices similarly indicate lower concern. These rationales indicate the need for at least one print copy of well-digitized digitally preserved text-only materials to be available for at least 20 years.

In order to guard against losses over time and assure the availability of a single copy after the stated time horizon, a greater number of print copies of any digitized title need to be secured today. In the exemplar scenario, a minimum of two page-verified print repository copies would be needed. When such well-digitized digitally-preserved text-only journals are held in two page-verified print repository environments, therefore, other libraries can safely withdraw their print holdings if they so
choose.

In candor, I haven’t read — nor am I likely to read — beyond the executive summary.  But it strikes me that these rationales mostly apply to archival libraries — the Library of Congress being the most obvious example — rather than to working university libraries.  There’s simply no reason for the average PhD-granting institution — let alone regional universities — to budget for print editions of journals, which are 1) ridiculously expensive, 2) take up enormous space, 3) can be accessed by only one user at a time, and 4) easily damaged through use or malice.  (Pages being ripped out are not uncommon.)  Most scholars will simply conduct their research from the comfort of their office computer, which has become increasingly the norm over the past 15 or so years.